JP2001196937A - Data coding method, data decoding method, and computer- readable recording medium recording program to allow computer to execute the method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To efficiently compress pseudo medium tone image data generated by error spread processing in soft ware processing by a microprocessor. SOLUTION: Any of states S0-S3 is selected as a state value S on the basis of a consecutive 0-bit number X, a consecutive 1-bit number Y, a value A just before the X, and a value B just before the Y, a value just before the S is selected as a T, and on the incidence of a dissident event, a state code Q denoting the state value S is generated. On the incidence of a coincident event in the case of X≠A, a numeral code P denoting the value X is generated, and in the case of Y≠B, a numeral code P denoting the value Y is generated, and on the incidence of a regression event, a numeral code P denoting a value of a reparative number N is generated prior to the production of the state code.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention
A data encoding method of sequentially inputting a set with the number of bit continuations as an input unit, converting the data into a predetermined code string and outputting the code string, inputting a code string obtained by scanning code string data, and inputting a bit continuation number of 0 The present invention relates to a data decoding method for restoring and outputting a run-length numerical sequence in units of a set of X and a bit continuous number Y of 1 and a computer-readable recording medium storing a program for causing a computer to execute the method.

[0002]

2. Description of the Related Art Conventionally, in a printing apparatus such as a laser printer, image data to be printed is compressed (encoded) and temporarily stored in a memory, and the compressed data stored in the memory is decompressed (decoded). Meanwhile, a memory saving technique for reducing the memory capacity of a memory for storing image data by performing actual printing is known.

The original purpose of such a memory-saving technique is to reduce the memory capacity of a memory for storing image data to reduce costs. Therefore, complicated dedicated hardware for compressing or expanding image data is used. It is not appropriate to provide them, and it is necessary to cause the microprocessor of the printing apparatus to perform these processes by software.

Here, as a coding technique for a two-dimensional still image, a two-dimensional compression technique for performing data compression in consideration of data correlation in the sub-scanning direction or a predictive coding technique for performing predictive coding using an arithmetic code is used. Is known as a technique capable of obtaining a high compression ratio.

However, if these encoding techniques are implemented in software using a microprocessor of a printing apparatus, processing delays occur, and in recent years, the printing resolution and the printing speed are improved day by day. Using coding techniques is not efficient. Therefore, when a microprocessor performs compression and decompression, a relatively simple compression technique called one-dimensional compression is widely used.

Conventional one-dimensional compression techniques include MH used in run-length processing and facsimile machines.
(Modified Haffman) codes are widely known, but in the case of image data in which the color (black and white) frequently changes in a short cycle, such as a pseudo halftone image, it cannot be efficiently compressed. In particular, since the opportunity to print image data has increased due to the development of personal computers and networks, it is not efficient to adopt this MH code as a memory saving mechanism.

For this reason, in order to improve the compression ratio of image data in which colors frequently change in a short cycle, an encoding technique for detecting the repeatability of image data and run length is known. For example, Japanese Patent No. 2683506 (Prior Art 1) discloses that a control word capable of describing both the length of a white word and the length of a black word is used so that the white word and the black word are regulated at a relatively short cycle. A data compression (expansion) method and apparatus for efficiently compressing image data that appears in a random manner is disclosed.

Specifically, in the prior art 1, a certain word length (byte) is defined, data units that are all white are treated as white words, and those that are not all white are treated as black words. Since the number of occurrences, the contents of black words, control words, etc. are encoded in word length units,
Since a special code table is not required for encoding and decoding, and processing can be performed in word length units, sufficient processing performance can be expected even if processing is performed by a microprocessor.

Japanese Unexamined Patent Publication No. Hei 9-65147 (Prior Art 2) stores run lengths of white and black, and a case where a run length value of a certain color matches a run length value immediately before the same color. Discloses an image signal compression (decompression) method and apparatus for improving a compression ratio and a processing speed by generating a predetermined repetition code.

[0010]

However, in the above-mentioned prior art 1, since the repeatability is detected in word string units, the repetition period included in the target image data and the word length are equal to each other. There is a problem that only a high compression ratio can be obtained only when it is a multiple, and sufficient compression cannot be obtained for general image data.

[0011] Further, in the above-mentioned prior art 2, when the repetition code generated when the run lengths match does not match, the data amount is sufficiently smaller than the code word to be coded when the repetition code does not match. There is a problem that only the compression ratio is improved. In other words, assigning shorter codewords to repetition codes lengthens codewords with run lengths that do not match relatively, so that a sufficient compression ratio is applied to image data in which colors frequently change in a short cycle. It is difficult to obtain a code set having

Specifically, in the embodiment of the prior art 2, the code set is similar to the MH code, and the processing by the microprocessor cannot be expected to be as fast as the run-length code technique. . In addition, since a storage unit for storing run-length values is provided for each color, it is necessary to perform a process of selecting a reference / storage destination according to a color when comparing and storing run-length values. There is also a problem that it hinders high-speed processing by the processor.

In the two-dimensional compression technique, the predictive coding technique and the MH coding, a line buffer,
There is a problem that a relatively large capacity memory such as a probability table and a code table is required. For example, 1
200 dpi (dot / inch) resolution, 25 ppm
If the printing speed is (pages / minute), the pixel frequency is 10
Generally, it exceeds 0 MHz. It is extremely difficult to operate an encoding / decoding hardware circuit including the above-mentioned memory at such a frequency, and even if the circuit is operated in parallel and operated at a low frequency, the parallel operation including the table is performed. Therefore, a large-scale circuit is required, which results in an increase in cost.

Japanese Patent Application No. Hei 11-12479 discloses that, when events in which the run length value coincides with the immediately preceding run length value of the same color continue, the number of codes is reduced by encoding the continuous number. A data encoding and decoding method is described in which a high compression rate and a high processing speed can be achieved at the same time even for pseudo halftone image data. In the case of image data including a pseudo halftone image, a sufficient compression ratio cannot be obtained.

According to the present invention, pseudo-halftone image data generated by an error diffusion process can be efficiently compressed in order to solve the above-described problems of the prior art, and is compatible with software processing by a microprocessor. An object of the present invention is to provide a data encoding method, a data decoding method, and a computer-readable recording medium recording a program for causing a computer to execute the method.

[0016]

Means for Solving the Problems To solve the above-mentioned problems,
In order to achieve the object, a data encoding method according to the present invention is characterized in that a run-length numerical sequence obtained by scanning data is sequentially set using a set of a 0-bit continuous number and a 1-bit continuous number as an input unit. A data encoding method for inputting, converting into a predetermined code string, and outputting the same, wherein the number of consecutive 0 bits, the number of consecutive 1 bits, the number of consecutive 0 bits immediately before, and the number of consecutive 1 bits A calculating step of calculating a state value corresponding to the data of the input unit based on the number of continuous bits; and a state value corresponding to the data of the input unit based on the state value calculated by the calculating step and the immediately preceding state value. Generating a state code and a numerical code.

According to the first aspect of the present invention, the input unit is determined based on the number of consecutive 0 bits, the number of consecutive 1 bits, the number of consecutive 0 bits immediately before, and the number of consecutive 1 bits immediately before. The state value corresponding to the data is calculated, and the state code and the numerical code corresponding to the data of the input unit are generated based on the calculated state value and the immediately preceding state value. The pseudo halftone image data can also be efficiently compressed.

According to a second aspect of the present invention, in the data encoding method according to the first aspect of the present invention, the calculating step comprises the step of calculating the number of consecutive 0 bits X, the number of consecutive 1 bits Y,
When the number of consecutive bits A of the immediately preceding 0 bits and the number of consecutive bits B of the immediately preceding 1 bit are in a relationship of X ≠ A and Y ≠ B, the value of the state S0 is set to the state value S, and X = A and Y ≠ B, the value of the state S1 is defined as the state value S, and X ≠ A and Y =
B, the value of the state S2 is set to the state value S, and X
= A and Y = B, and the value of the state S3 is set as the state value S.

According to the second aspect of the present invention, the continuous number X of 0 bits, the continuous number Y of 1 bits, the continuous number A of immediately preceding 0 bits, and the continuous number B of immediately preceding 1 bits are: X ≠
When the relationship is A and Y ≠ B, the value of the state S0 is set to the state value S. When the relationship X = A and Y ≠ B, the state S1 is set.
Is the state value S, the value of the state S2 is the state value S when X ≠ A and Y = B, and the value of the state S3 is the state when X = A and Y = B. Since the value S is set, the data can be efficiently compressed using the four state values.

According to a third aspect of the present invention, in the data encoding method according to the second aspect of the present invention, in the generating step, the immediately preceding state value T is the state S3, and When a coincidence event equal to the state value T is considered, this is regarded as a repetition event, and a repetition number calculation step of calculating a repetition number N in which the repetition event occurs continuously; and A regression event determining step of determining a regression event when the status value S and the immediately preceding status value T differ from each other or when input is completed; and a status indicating the value of the status value S when the mismatch event occurs. A state code generation step of generating a code, and when the generation of the state code by the state code generation step ends or the coincidence event occurs,
A first numerical code generating step of generating a numerical code indicating a value of the number of consecutive bits X of 0 or the number of consecutive bits Y of 1 bit, and the state code generating step when the regression event occurs. A step of generating a numerical code indicating a value of the number of iterations N calculated in the number of iterations calculating step, prior to generating the state code.

According to the third aspect of the present invention, when the immediately preceding state value T is the state S3 and the state value S is equal to the immediately preceding state value T, a coincidence event is defined as a repeated event. Considering this, the number of repetitions N at which the repetitive event occurs continuously is calculated, and after the repetitive event occurs, the state value S and the immediately preceding state value T
Is determined to be a regression event when a mismatch event or input completion is different from the above, and a state value S is determined when a mismatch event occurs.
Is generated, and when the generation of this status code is completed or a coincidence event occurs, a numerical code indicating the value of the number of consecutive bits X of 0 or the number of consecutive bits Y of 1 is changed to When a regression event occurs, a numerical code indicating the value of the number of repetitions N is to be generated prior to the generation of the status code, so that the concept of a coincidence event, a non-coincidence event, a repetition event, and a regression event is used. Data can be efficiently compressed.

According to a fourth aspect of the present invention, in the data encoding method according to the third aspect of the present invention, when the encoding is started, each of T, A, B and the number of repetitions N is set to a predetermined value. An initializing step of storing an initial value of, an inputting step of inputting a predetermined run-length value to the X and Y, and a state value generating step of generating the state value S after the input by the inputting step. The generation step includes a coincidence event determination step of determining occurrence of the coincidence event or the non-coincidence event in response to the state value S generated by the state value generation step; and the repetition event in response to the occurrence of the coincidence event. A repetition event determining step of determining whether or not the occurrence of a repetition event, a repetition number adding step of adding the repetition number in response to the occurrence of the repetition event, and in response to the occurrence of the mismatch event or the end of the encoding target data Said times A regression event determination step of determining whether an event has occurred; a repetition number code generating step of generating a repetition number code indicating the value of the repetition number N in response to the occurrence of the regression event; A regression step of returning the number of iterations N to a predetermined initial value after code generation, and a state code representing the value of the state value S in response to the occurrence of the mismatch event or the initialization of the number of iterations N by the regression step After the state code generation step of generating the state code and the determination of the occurrence of the repetition event by the repetition event determination step, when the X does not coincide with A, a numerical code corresponding to the X And a numerical code generating step of generating a numerical code corresponding to Y when the Y does not match B, and generating a numerical code by the numerical code generating step or repeating by the repeating number adding step After the addition of N,
A value updating step of updating the content of A to X, updating the content of B to Y, and updating the content of T to S.

According to the fourth aspect of the present invention, when encoding is started, a predetermined initial value is stored in each of T, A, B and the number of repetitions N, and a predetermined run length value is stored in X and Y. Input, generate a state value S after this input, determine whether a matching event or a non-matching event has occurred according to the generated state value S, and determine whether or not a repetitive event has occurred in response to the occurrence of the matching event In response to the occurrence of a recurring event, the number of repetitions is added, and in response to the occurrence of a discrepancy event or the end of the data to be encoded, it is determined whether or not a regression event has occurred. A repetition number numerical code indicating the value of the repetition number N is generated, and after the repetition number code is generated, the repetition number N is returned to a predetermined initial value, and a state value is generated in response to occurrence of a mismatch event or initialization of the repetition number N. A state code representing the value of S is generated, and a state code is generated or After the determination of the occurrence of repetitive events, X is A
Generates a numerical code corresponding to the X when not equal to
If Y does not match B, generate a numerical code corresponding to Y, and after generating the numerical code or adding the number of repetitions N, update the contents of A to X, update the contents of B to Y, Is updated to S, so that data compression suitable for software processing by the microprocessor can be performed.

According to a fifth aspect of the present invention, in the data decoding method, a code string obtained by scanning code string data is input, and a set of a 0-bit continuation number X and a 1-bit continuation number Y is formed. What is claimed is: 1. A data decoding method for restoring and outputting a run-length numerical sequence as a unit, comprising: a storage unit S for storing a restored state value when decoding is started; An initialization step of substituting a predetermined initial value into a storage unit C for storing a value corresponding to X and a storage unit D for storing a value corresponding to Y among the run length values to be restored, and An input code type determining step of inputting the code unit of (i) and determining whether the code is the numerical code P or the status code Q. If the input code type determining step determines that the code is a numerical code, To be restored to a numerical value and stored in the storage unit L A state code restoring step, a state code restoring step of restoring the code to the state value when it is determined by the input code type judging step to be the state value and storing the same in the S, and a restored numerical value stored in the L. The state value S at the time of the state S0, X = A in the relation of X ≠ A and Y ≠ B
And a state S1 in a relation of Y ≠ B, a state S2 in a relation of X ≠ A and Y = B, or a state S3 in a relation of X = A and Y = B (where A is a continuation of the immediately preceding bit of 0). A number, and B is the number of continuations of the immediately preceding 1 bit), and when a new state value is stored in S, the value of S is After the value of L is stored in the C when the value of S is determined to be S0 in the S3 state determination step of determining whether or not they match, and in the identification step, the next numerical code Inputting and restoring, and storing the value of L in the D when the value of S is identified as S1 in the first run value updating step of storing the value in the D and storing the value of S in the identification step The value of S is S2 by the second run value updating step to be performed and the identification step. If identified, a third run value update step of storing the value of L in the C, and a point in time when the update in the first, second or third run value update step is completed or the S3 state determination When it is determined in the step that the value of S matches the state value S3, the values of C and D are determined.
And a first output step of outputting the values of C and D as a set of run-length values. If the value of the S value is identified as S3 by the identification step, the values of C and D are A second output step of outputting a run length value the number of times corresponding to the value of L.

According to the fifth aspect of the present invention, when decoding is started, the storage unit S for storing the restored state value and the value corresponding to the X among the restored run length values are stored. A predetermined initial value is substituted into a storage unit C and a storage unit D that stores a value corresponding to the Y among the run length values to be restored and initialized, a predetermined code unit is input, and the code is a numerical code P Or the status code Q, and if it is determined that the code is a numerical code, the code is restored to a numerical value and stored in the storage unit L. The state value S at the time when the restored numerical value is stored in L is stored in S, and the state value S at the time when the restored numerical value is stored in L is the state S0 in the relation of X 関係 A and Y ≠ B, State S in relation
1, a state S2 or X in which X ≠ A and Y = B
= A and Y = B are identified, and when a new state value is stored in S,
Is determined whether the value of S is equal to S3. If the value of S is identified as S0, the value of L is stored in C, the next numerical code is input and restored, and stored in D Then, when the value of S is identified as S1, the value of L is stored in D,
When it is determined that the value of S is S2, the value of L is stored in C, and when the update is completed or the value of S is the state value S3
When it is determined that the values match, the values of C and D are output as a set of run-length values. If the value of S is identified as S3, the values of C and D are Since the number of times corresponding to the value of L is output as the length value, the compressed halftone image data generated by the error diffusion process can be efficiently restored.

Further, the recording medium according to the invention of claim 6 is:
By recording a program for causing a computer to execute the method according to any one of claims 1 to 5, the program becomes machine-readable, thereby
The operation of any one of claims 1 to 5 can be realized by a computer.

[0027]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS With reference to the accompanying drawings, a data encoding method and a data decoding method according to the present invention and a computer-readable recording medium storing a program for causing a computer to execute the method will be described below. Embodiments will be described in detail. In the present embodiment,
It is assumed that pseudo halftone image data generated by the error diffusion process is encoded or decoded.

First, the concept of a data encoding method and a data decoding method according to the present embodiment will be described. FIG. 1 is an explanatory diagram for explaining the concept of a data encoding method and a data decoding method according to the present embodiment.

In the figure, in a run-length numerical sequence obtained by scanning data, the number of consecutive bits (hereinafter referred to as “X”) having a value of “0” and the number of bits having a value of “1” are shown. Is sequentially input as an input unit, and is converted into a predetermined code string and output.

Specifically, the value immediately before X is "A",
The value immediately before Y is “B”, the state of X ≠ A and Y ≠ B is S0, the state of X = A and Y ≠ B is S1, the state of X ≠ A and Y = B is S2, X = A
The state of Y = B is S3.

In addition, a state value having one of S0 to S3 according to this state is defined as “S”, and a value immediately before S is defined as “T”. An event where S = T is defined as a “general event”, an event where S な る T is defined as a “mismatch event”, and if a coincidence event occurs when T is S3, it is defined as a “repeated event”. The number of consecutive occurrences of an event is referred to as “the number of repetitions N”.

Further, after the occurrence of the repetitive event, if a mismatch event occurs or the input is completed, the state is defined as a "regression event". If a mismatch event occurs, a status code Q indicating the value of the status value S is displayed. At the time when the code generation is completed or when a coincidence event occurs, a numerical code P indicating the value of X is generated when X ≠ A, and the value of Y is generated when Y ≠ B. When a regression event occurs, a numerical code P indicating the value of the number of repetitions N is generated prior to generation of the state code.

FIG. 1 shows a case where the data to be compressed starts with 3-bit “0” and continues with 4-bit “1” and 3-bit “0”. When the run length measurement is started from “0”, the numerical sequence of X and Y obtained by scanning the data is a continuous number of “0” and a continuous number of “1”, respectively. Are sequentially input to the encoding unit.

In the figure, each numerical value, event, and generated code for each input value are shown. For each event, “○” is described when there is an event, and when there is no event, Is blank. Also, for each code,
If there is a code output, “○” is described, and if no code is output, it is blank.

Here, although there are two attributes, a state code and a numerical code, the numerical codes are shown in three types so that it is possible to grasp which numerical value corresponds to which numerical code. Strictly speaking, there is no immediately preceding value (A, B and T) in the first X, Y and S of the data, but here the initial values of A, B and S are all “0”. And

As described above, in the present embodiment, when events in which the run length value coincides with the run length value immediately before the same measurement target value continue, the number of codes is reduced by encoding the continuous number. In this case, the number of codes is efficiently reduced by inheriting the feature of the related art of performing the above by generating a numerical code P corresponding to the number of repetitions N when S3 is repeated a plurality of times. In addition, when one of X or Y matches the previous value, the other does not match, that is, S1
Alternatively, even when the state of S2 is continuous, the number of codes is efficiently reduced.

In particular, in the error-diffused image data, there is a high probability that the states of S1 and S2 are continuous.
The effect of reducing the number of codes according to this embodiment is great. For example, in the figure, although 12 sets, that is, 24 run-length values are input, the number of output codes is reduced to 11, and it can be seen that the number of codes is reduced.

Next, an encoding procedure used in the present embodiment will be described. FIG. 2 is a flowchart showing the encoding procedure used in the present embodiment, and FIG. 3 is a flowchart showing the non-output N encoding procedure shown in steps S209 and S219 in FIG.

As shown in FIG. 2, first, an initialization process for setting a predetermined constant C to A, B and T and setting 0 to N is performed (step S201). Here, for convenience of explanation, the same constant C is set for A, B and T, but different values can be set. However, it is desirable that the initial value of T be the value shown in S0 in order to reduce the number of codes. The reason is that the state value S becomes S0 with respect to the first input of X and Y at the highest probability, so that the generation of a state code due to the occurrence of a meaningless mismatch event can be suppressed. Further, it is desirable to set the initial values of A and B to values different from the values at the start of the run length measurement in order to reduce the number of codes.

When the initialization process is completed, an input process for inputting and storing a set of run-length values in X and Y is performed (step S202), and a predetermined state S corresponding to the input is executed. Is performed (steps S203 to S207).

More specifically, after initializing by substituting 0 for the S value (step S203), the X value and the A value are compared (step S204). (S204 affirmative) '1' is added to this S (step S205), and if they do not match (step S2)
04 No) The process directly proceeds to step S206. Further, the Y value is compared with the B value (step S206), and if they match (Yes at step S206), the process proceeds to step S206.
2 'is added (step S207). Therefore, S
The value representing 0 is 0, the value representing S1 is 1, the value representing S2 is 2, and the value representing S3 is 3.

Thereafter, a matching event determination process is performed to determine whether or not the generated state value S matches T (step S208).
(8 affirmative), it is determined that the event is a coincidence event, and if they do not match (No at Step S208), it is determined that the event is a non-coincidence event.

More specifically, if a coincidence event has occurred (Yes at step S208), it is determined whether or not the S value is 3 (step S211).
If (step S211 is affirmative), it is determined that there is a repetition event, the value of the number of repetitions N is incremented (step S212), and the routine goes to step S217.

On the other hand, if a mismatch event has occurred (No at step S208), it is determined that no repetition event has occurred, and non-output N encoding is performed (step S209). Specifically, as shown in FIG.
Is determined whether or not the value of is larger than 0 (step S30).
1) If it is larger than 0 (Yes at step S301), it is determined that a regression event has occurred, and numerical encoding (N) is performed to output a numerical code (step S301).
302), and initialize by substituting 0 for N (step S3)
03). On the other hand, if the value of the number of iterations N remains 0 (No at Step S301), it is determined that no regression event has occurred.

Then, when the non-output N encoding is completed, state encoding (S) for generating and outputting a state code representing the value of the state S is performed (step S210), and a run numerical code output process is performed. (Steps S213 to S2)
16).

More specifically, first, the X value and the A value are compared (step S213), and if they do not match (No at step S213), a state code representing the value of X is generated and output. (X) (Step S21)
4). Further, the Y value and the B value are compared (step S21).
5) If they do not match (No at Step S215), a numerical code (Y) for generating and outputting a state code representing the value of Y is performed (Step S216).

Thereafter, A is updated with the X value, B is updated with the Y value, and the value immediately before T is updated with the S value is updated (step S217). If there is data to be encoded (step S218) No) The process proceeds to step S202 and the above process is repeated. If there is no more data to be encoded (Yes at Step S218), the already-outputted N-code is performed (Step S219), and the process ends.

By performing the above-described series of processing, when a mismatch event occurs, a state code Q indicating the value of the state value S is generated. When the code generation is completed or when a match event occurs, X is generated. If ≠ A, a numerical sign P indicating the value of X is generated, if Y ≠ B, a numerical sign P indicating the value of Y is generated, and if a regression event occurs,
Prior to generation of the state code, a numerical code P indicating the value of the number of repetitions N can be generated and encoded.

When the above procedure is realized by software, appropriate memories may be allocated to X, Y, A, B, S, T and N, but general-purpose registers of a microprocessor are used. It is desirable to improve processing performance. The constant C and each control can be realized by describing them as a program, that is, an instruction code.

In the above series of description, for convenience of explanation,
Run-length measurement input (step S202) and data end determination related thereto (step S218) and code generation (steps S210, S214, S216, S)
The detailed description of 302) has been omitted.

Next, a data decoding procedure according to the present embodiment will be described. FIG. 4 is a flowchart showing a data decoding procedure according to the present embodiment. As shown in the figure, when the decoding process is started, S which stores a restored state value, A which stores a value corresponding to X among the restored run-length values, and a value which corresponds to Y Is initialized by substituting a constant C into B and storing B, respectively (step S
401). In order for decryption to be performed correctly,
The initial values to be assigned to A, B, and S need to be the same as the values assigned to A, B, and S during the initialization process at the time of encoding (step S201 in FIG. 2).

Then, after completing this initialization processing,
The code input is substituted for K (step S402), and an input code determination process (step S403) for determining whether this K is the numerical code P or the status code Q is performed. As a result, if it is determined that the code is a numerical code (step S
(403 affirmative), a numerical value restoring process (step S404) of restoring the code stored in K to a numerical value and storing it in L is performed, and the process proceeds to a state identification process (steps S405 to 407).

Specifically, it is confirmed whether or not the state value S is 0 (step S405). If the state value S is 0 (Yes at step S405), the process corresponding to S0, ie, A The value of L, which is the restored run-length value, is stored (step S408), and the next code unit is input and K
(Step S409), the code stored in K is restored to a numerical value and stored in L (Step S410), the value of L is stored in B (Step S411), and A is set as the output X. B is output as a set of run-length values with output Y (step S414).

If the status value S is not 0 (No at Step S405), it is checked whether the status value S is 1 (Step S406), and if the status value S is 1 (Step S406). (S406 affirmative), a process corresponding to S1, that is, a process of storing the L value in B is performed (step S41).
2) Output a set of run-length values where A is the output X and B is the output Y (step S414).

Further, if the state value S is not 1 (No at Step S406), it is checked whether or not the state value S is 2 (Step S407). If the state value S is 2 (Step S407). (Affirmative in S407), a process corresponding to S2, that is, a process of storing the L value in A is performed (step S4).
13) Output a set of run-length values where A is the output X and B is the output Y (step S414).

If the status value S is not 2 (No at Step S407), a process corresponding to S3 is performed.
Specifically, after the value of L is stored in N as an initial value (step S417), it is output as a set of run-length values in which A is output X and B is output Y (step S41).
8) After decrementing N (step S419),
If the value of N is not 0 (No at Step S420),
The process moves to step S418 and the process is repeated. And
When the value of N becomes 0 (Yes at step S420),
An end determination is made (step S421), and if not (step S421 negative), the above-described step S4
Move to 02.

If it is determined in step S403 that the code is a numerical code (No in step S403),
The code stored in K is restored to a state value and stored in S (step S415), and it is confirmed whether or not S is 3 (step S416). As a result, when S is 3 (Yes at Step S416), the process proceeds to Step S414, and when S is not 3 (No at Step S416), the process proceeds to Step S421.

By performing the above series of processing, when a state code is input and restored to a predetermined state value S, if this S is in the state of S3, a set of run-length values is omitted. Therefore, a set of run-length values stored in A and B at that time can be output without waiting for the input of a numerical code.

When a numerical code is input, restored to a predetermined numerical value and stored in L, if this S value is S0, a set of run-length values is encoded into a set of numerical codes. Therefore, a set of run-length values is output using the value stored in L and the value obtained by decoding the next numerical code.

When the S value is S1, 0
Is omitted, a set of run-length values is output using the value stored in A and the restored numerical value L. If the S value is S2, 1 Is omitted, the output of a set of run-length values is performed using the value stored in B and the restored numerical value L.

Further, when the S value is S3,
Since the restored number is the number of iterations N and the N sets of run-length values are omitted, the output operation of one set of run-length values is performed using the values stored in A and B at that time. Repeat for the number of times. For this reason, the original run-length sequence can be decoded from the code sequence data generated by the data encoding procedure shown in FIG.

When the above procedure is implemented by software, appropriate memories may be allocated to X, Y, A, B, S, T, K, L and N. It is desirable to use a general-purpose register in order to improve processing performance. The constant C and each control can be realized by describing them as a program, that is, an instruction code.

Some of the above series of processing can be omitted. For example, N for counting and storing the number of repetitions and L for storing the restored numerical value can be omitted. Can also. One of the X and Y for storing the output run length value can be substituted for the other.

Further, in the above series of descriptions, run-length numerical output (steps S414 and S414) is provided for convenience of explanation.
418) and the related data end determination (step S421) have been omitted.

Next, a code set used in the data encoding process and the data decoding process shown in FIGS. 2 and 4 will be described. FIG. 5 is an explanatory diagram illustrating an example of a code set used for the data encoding process and the data decoding process illustrated in FIGS. 2 and 4.

As shown in the figure, the code set used in the present embodiment is composed of first to sixth types of code units. Specifically, the first code format P1 to the third code The format P3 is a numerical code P indicating a numerical value, and a fourth code format Q
The first to sixth code formats Q3 are status codes Q indicating status values.

Here, the first code format P1 has a 4-bit width and is identified using the fact that the 4-bit numerical value is less than “10”, and 1 is assigned to the numerical value. The added value corresponds to the numerical value L. The second code format P2 has an 8-bit width, is identified by the value of the preceding upper 4 bits in the code string being “11” or “12”, and the value of the lower bit is “2”. The value obtained by adding 11 ″ corresponds to the numerical value L.

The third code format P3 has a bit width of at least an 8-bit width and a multiple of 4, and is identified by the value of the preceding upper 4 bits in the code string being "12". As for these 4 bits, the value obtained by adding "43" to the numerical value obtained by concatenating (joining) the lower 3 bits of each 4 bits upward until a numerical value of "8" or more is detected corresponds to the numerical value L. I do.

The fourth code format Q1 has a 4-bit width and is identified by the fact that the 4-bit value is "13", and the lower 2 bits of the value obtained by adding "1" to the immediately preceding status value. Corresponds to the new state value. The fifth code format Q2 is
It has a 4-bit width, and is identified by the fact that the numerical value of the 4-bit is “14”, and the lower 2 bits of the value obtained by adding “2” to the immediately preceding status value correspond to the new status value. The sixth code format Q3 has a 4-bit width, and is identified by the fact that the numerical value of the 4-bit is “15”, and the lower 2 bits of the value obtained by adding “3” to the immediately preceding status value are new. Corresponds to the state value.

Although the state value can take four values, the state code is generated only when a change occurs in the state value. The transition target from one state value to another state value is 3
Since there are only three, the number of state codes is reduced to three.

In the numerical code P, a code whose value is “0” is not explicitly shown, but it is desirable to represent “0” by using a conversion procedure described later. However, it should be noted that this code format does not restrict the expressible value range in principle, and an arbitrary integer can be expressed when the third code format P3 is used. This is because the temporary storage used for the process of restoring a numerical value has a finite bit width, but since the length of the P3 code is not restricted, the processing result by the concatenation and addition may be “0” due to a carry, or “0”. This is because a sign that is a negative number can be generated.

Next, when the code set shown in FIG. 5 is used, input code type determination, state coding control, state restoration processing,
The numerical value encoding process and the numerical value restoration process will be described.
First, in the determination of the input code type (corresponding to step S403 in FIG. 4), if the value of the preceding upper 4 bits of the code is a value equal to or less than "12", it is determined that the code is a numerical code.
Otherwise, it can be determined that it is a status code.

FIG. 6 is a flowchart showing a state encoding control procedure when the code set shown in FIG. 5 is used. FIG. 7 is a state decoding control procedure when the code set shown in FIG. 5 is used. It is a flowchart which shows a procedure. In addition,
“K (n) ← numerical value” shown in the figure is the lower n of the numerical value on the right side.
Bits are output as codes, and "W" is
This is a working memory used for calculation.

As shown in FIG. 6, when state encoding is performed, a value obtained by subtracting the immediately preceding state value T from the state value S is obtained by adding
3 'is ORed bit by bit, and a value obtained by adding "12" to this is stored in the working memory W (step S601).
The lower 4 bits of the working memory W are output as a 4-bit code K (step S602).

As shown in FIG. 7, when state decoding is performed, a value obtained by subtracting "12" from the code value in response to the occurrence of one of the codes Q1 to Q3 is set to the state up to that time. The value is added to the value S, the logical product of the result value and the constant value '3' is calculated for each bit, and this is stored as a new state value in the state value S (step S701).

By performing the control shown in FIGS. 6 and 7, it is possible to generate a predetermined state code and restore a state value using the state code. For example, if the state value T immediately before encoding is “0” and the new state value S is “3”, “ST” becomes “3”, so this value and the constant value “3” The logical product of 'for each bit is'3'. Then, "15" obtained by adding "12" to this value, that is, Q3
A code is generated. On the other hand, when restoring the state code,
"K-12" becomes "3", and when a logical AND for each bit of a value "3" obtained by adding the previous state value "0" to this value and a constant value "3" is obtained, "3" is newly obtained. It can be seen that the state value S is properly restored.

If the state value T immediately before encoding is “3” and the new state value S is “0”, “ST”
Is "-3", and the logical product of this value and the constant value "3" for each bit is "1", so "13", which is obtained by adding "12" to this value, that is, a Q1 code is generated. On the other hand, when restoring the state code, “K-12” becomes “1”,
When a logical AND for each bit of a value '4' obtained by adding the previous state value '3' to this value and a constant value '3' is taken, '0' becomes a new state value S and is correctly restored. You can see that.

FIG. 8 is a flowchart showing a numerical value encoding control procedure when the code set shown in FIG. 5 is used, and FIG. 9 is a numerical decoding control procedure when the code set shown in FIG. 5 is used. It is a flowchart which shows a procedure. In addition,
“K (n) ← numerical value” shown in the figure is the lower n of the numerical value on the right side.
Bits are output as codes, and "W" is
This is a working memory used for calculation.

As shown in FIG. 8, here, the numerical value "V" to be encoded is converted into one of the codes P1 to P3 and output. As already mentioned, since the sign whose value is “0” is not defined in this numerical code P, here, when “V = 0”, this is converted into a constant value “Z”. 0 value conversion processing is performed. That is, since the frequency of V = 0 is very low, the compression rate is improved by assigning this to a relatively long codeword. Specifically, steps S801 to S801 shown below
Step S804 corresponds to this 0-value conversion processing, and step S80
Steps 5 and after correspond to a process of generating a codeword from a numerical value.

As shown in the figure, first, it is determined whether or not the encoding target value V is equal to or larger than a constant value Z (step S801).
Affirmative) Increment the value of V (step S80)
2). If it is not equal to or greater than Z (step S801)
(No) It is determined whether V is 0 (step S80).
3) If Z is 0 (Yes at Step S803), the constant value Z is stored in V (Step S804).

When the zero-value conversion processing is completed, it is determined whether or not the current encoding target value V is equal to or less than "10" (step S805). (Yes at step S805), V is set to “1”.
Is output as a code (step S806), and the numerical value encoding ends.

On the other hand, if V is not equal to or less than "10" (No at step S805), it is determined whether or not this V is equal to or less than "42" (step S807). (Yes at Step S807) This V is set to “1”.
The lower 8 bits of the value obtained by adding "49" are output as a code (step S808), and the numerical encoding is terminated. From the definition of the P2 code, "K (8) ← 160 + (V-11)", that is, "K ( 8) ← V + 149 ”, so that the code generated here is nothing but the P2 code. This "
When 160 ″ is expressed in a binary number, it becomes “10100000”, so that the upper 3 bits are a code-specific constant.

In step S807, V
Is not less than or equal to "42" (No at Step S807), first, as the start identification code of the P3 code, "
A 4-bit code having a value of "12" is output (step S809), and a value obtained by subtracting "43" from V is stored in the working memory W (step S810), and the value of this W becomes "8". It is checked whether it is less than (Step S81)
1).

As a result, if the value is equal to or more than '8' (No at Step S811), the value of the lower 3 bits of the W value is output as a 4-bit code (Step S81).
2) After shifting the W value to the right by 3 bits and discarding the portion output as a code (step S813), the process returns to step S811.

On the other hand, if it is smaller than '8' (Yes at step S811), a value obtained by adding '8' to the W value is output as a 4-bit code (step S814).
End the numerical encoding. In addition, since the constant '7' is expressed as “0111” in binary notation, the code generated in step S812 is always a numerical value equal to or more than '8', and it can be clearly indicated that this code is the last of the numerical part. It can be seen that this code is nothing but the P3 code.

As shown in FIG. 9, when performing a numerical restoration, a code string coded in any of the code formats P1 to P3 is sequentially read in 4-bit units, and
Will be restored. “K ← Sign input”
The notation "" means that a 4-bit code is input to K as a numerical value. Also, here, “0-value inverse conversion processing”, which is the inverse conversion of the 0-value conversion processing shown in FIG. 8, is performed. Steps S901 to S911 described below correspond to a numerical value restoration process, and steps S912 and subsequent steps correspond to a 0-value inverse conversion process.

As shown in the figure, when performing the numerical value restoration processing, first, the first 4 bits of the code are input to K (step S901), and it is determined whether or not this K value is less than “10”. A determination is made (step S902). If the K value is less than “10” (Yes at Step S902), a value obtained by adding “1” to the K value is set as a restored numerical value.
(Step S903), and the process proceeds to a 0-value inverse conversion process described later. Note that a code whose first 4 bits are less than “10” is nothing but a P1 code.
Numerical values can be restored by adding 1 '.

On the other hand, if the K value is not less than “10” (No at Step S 902), the K value becomes “1”.
It is determined whether or not “1” is “2” (step S904).
If it is not 2 "(No at Step S904), the value obtained by shifting the K value left by 4 bits is stored in W, the next 4 bits of the code are input to K, and W and the new K value are added. After the value obtained by subtracting “149” from the calculated value is stored in W as a restored numerical value (step S905), the process proceeds to a 0-value inverse conversion process described later.

On the other hand, if the K value is "12" (Yes at step S904), since the code is nothing but a P2 code having an 8-bit length, the value of W is initialized to "0" (step S904). S906), the next 4 bits of the code are input to K (step S907), the K value is added to W (step S908), and it is determined whether or not the K value is less than '8' (step S906). S909).

As a result, when the K value is less than '8' (Yes at step S909), since the sign of the numerical part follows thereafter, after shifting the value of W to the left by 3 bits (step S910), The process moves to step S907.

On the other hand, if the K value is not less than “8” (No at step S909), “3” is added to W up to that time.
The value obtained by adding 5 "is stored in W again (step S
911), and shifts to the 0-value inverse conversion process. In the P3 code,
A value obtained by adding "43" to a numerical value obtained by connecting (joining) the lower 3 bits of each of the 4 bits upwards until a value of "8" or more is detected as the 4 bits corresponds to the restored numerical value. In addition, since “8” indicating the end of the numerical part is added to the value of K when a numerical value equal to or more than “8” is detected, “43-8” is “3-8” here.
5 "is added.

Thereafter, a 0-value inverse conversion process for restoring the 0-value conversion process is performed. Specifically, it is determined whether or not the restored numerical value W is a value larger than the constant Z (Step S).
912) If the value is larger than Z (Yes at step S912), a value obtained by subtracting '1' from W is stored in W (step S913), and the numerical value restoration process ends.

On the other hand, if the numerical value W is equal to or smaller than Z (No at Step S912), it is checked whether the restored numerical value W matches the constant Z (Step S914).
If they match (Yes at step S914), this W
Is substituted for '0' (step S915), and the numerical value restoration processing ends.

As described above, in the present embodiment,
Based on the continuous number X of 0 bits, the continuous number Y of 1 bit, the value A immediately before X, and the value B immediately before Y, any one of the states S0 to S3 is set as the state value, and When the immediately preceding value is T, a status code Q indicating the value of the status value S is generated when a mismatch event occurs, and X ≠ A at the time when code generation is completed or when a match event occurs. Generates a numerical code P indicating the value of X, generates a numerical code P indicating the value of Y when Y ≠ B, and iterates prior to generation of the state code when a regression event occurs. Since the numerical code P indicating the value of the number N is generated, the pseudo halftone image data generated by the error diffusion process can be efficiently compressed by software processing by the microprocessor.

[0095]

As described above, according to the first aspect of the present invention, the continuation number of 0 bits, the continuation number of 1 bit, the continuation number of the immediately preceding 0 bit, and the continuation of the immediately preceding 1 bit The state value corresponding to the data of the input unit is calculated based on the number, and based on the calculated state value and the immediately preceding state value,
Since the state code and the numerical code corresponding to the data of the input unit are configured to be generated, a data coding method capable of efficiently compressing the pseudo halftone image data generated by the error diffusion process is obtained. It works.

According to the second aspect of the present invention, the continuation number X of 0 bits, the continuation number Y of 1 bit, the continuation number A of the immediately preceding 0 bit, and the continuation number B of the immediately preceding 1 bit are obtained. , X
When the relationship is ≠ A and Y ≠ B, the value of the state S0 is set to the state value S, and when the relationship X = A and Y ≠ B, the state S is set.
1 is the state value S, and when X ≠ A and Y = B, the value of the state S2 is the state value S, and X = A and Y = B
Since the state S3 is set to the state value S in the case of the following relation, the effect of obtaining a data encoding method capable of efficiently compressing data using four state values is obtained. Play.

According to the third aspect of the present invention, when the immediately preceding state value T is the state S3, and the state value S is equal to the immediately preceding state value T, this event is repeated. And the number of repetitions N at which the repetitive event occurs continuously is calculated, and after the repetitive event occurs, when the state value S and the immediately preceding state value T are different events or the input is completed, a regression event is performed. When a mismatch event occurs, a status code indicating the value of the status value S is generated. When the generation of this status code ends or when a match event occurs, the number of consecutive bits X of 0 or Since a numerical code indicating the value of the number of consecutive 1 bits Y is generated, and when a regression event occurs, a numerical code indicating the value of the number of repetitions N is generated prior to generation of the state code. , Discrepancy event, repetition event,
There is an effect that a data encoding method capable of efficiently compressing data using the concept of the regression event is obtained.

According to the fourth aspect of the present invention, when encoding is started, a predetermined initial value is stored in each of T, A, B and the number of repetitions N, and a predetermined run length value is stored in X and Y. Is input, and a state value S is generated after the input. The occurrence of a matching event or a mismatching event is determined according to the generated state value S, and the presence or absence of a repetitive event is determined in response to the occurrence of the matching event. Judge, add the number of repetitions in response to the occurrence of the repetition event, determine whether a regression event has occurred in response to the occurrence of a mismatch event or the end of the data to be encoded, and respond to the occurrence of the regression event. To generate a numerical code for the number of repetitions indicating the value of the number of repetitions N. After the numerical code for the number of repetitions is generated, the number of repetitions N is returned to a predetermined initial value. Generate a state code representing the value of the value S;
After the state code is generated or the occurrence of the repetition event is determined, a numerical code corresponding to X is generated when X does not match A, and a numerical code corresponding to Y is generated when X does not match B. , After generating a numerical code or adding the number of iterations N,
Since the contents of A are updated to X, the contents of B are updated to Y, and the contents of T are updated to S, the data can be compressed so as to be suitable for software processing by the microprocessor. There is an effect that an encoding method can be obtained.

According to the fifth aspect of the present invention, when the decoding is started, the storage unit S for storing the restored state value,
A predetermined initial value is assigned to a storage unit C for storing a value corresponding to the X among the restored run-length values and a storage unit D for storing a value corresponding to the Y among the restored run-length values. Initialize, input a predetermined code unit and determine whether the code is a numerical code P or a status code Q,
If it is determined to be a numerical code, the code is restored to a numerical value and stored in the storage unit L. If it is determined to be a state code, the code is restored to a state value and stored to S, and restored to L When the value of the state value S at the time when the numerical value is stored is a state S0 in a relation of X ≠ A and Y ≠ B, a state S1 in a relation of X = A and Y ≠ B, X ≠ A and Y = B , Or the state S3 in the relationship of X = A and Y = B, and when a new state value is stored in S, the value of the S matches the value of S3. It is determined whether or not to perform
, When the value of L is stored in C, the next numerical code is input and restored and stored in D, and when the value of S is identified as S1, the value of L Is stored in D, and when the value of S is identified as S2, the value of L is stored in C, and when the update is completed or when it is determined that the value of S matches the state value S3, , C, and D are output as a set of run-length values, and if the value of S is identified as S3, the values of C and D correspond to the value of L as a set of run-length values. Since it is configured to output as many times as possible, an effect is obtained that a data decoding method capable of efficiently restoring the compressed halftone image data generated by the error diffusion process can be obtained efficiently.

According to the invention of claim 6, according to claim 1,
Recording a program that causes a computer to execute the method described in any one of (1) to (5), thereby making the program readable by a machine.
This provides an effect that a recording medium capable of realizing any one of the operations (1) to (5) by a computer is obtained.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram for explaining the concept of a data encoding method and a data decoding method according to an embodiment of the present invention.

FIG. 2 is a flowchart showing an encoding procedure used in the present embodiment.

FIG. 3 is a flowchart showing a non-output N encoding procedure shown in steps S209 and S219 of FIG. 2;

FIG. 4 is a flowchart showing a data decoding procedure according to the present embodiment.

FIG. 5 is an explanatory diagram showing an example of a code set used for the data encoding process and the data decoding process shown in FIGS. 2 and 4;

6 is a flowchart showing a state encoding control procedure when the code set shown in FIG. 5 is used.

FIG. 7 is a flowchart showing a state decoding control procedure when the code set shown in FIG. 5 is used.

8 is a flowchart showing a numerical encoding control procedure when the code set shown in FIG. 5 is used.

FIG. 9 is a flowchart showing a numerical value decoding control procedure when the code set shown in FIG. 5 is used.

[Explanation of symbols]

 X Number of continuations of Y 1 Number of continuations of 1 A X value immediately before B Y value immediately before S State value T S value immediately before N, P, P1, P2, P3 Numeric code Q, Q1, Q2, Q3 Status code

Claims (6)

[Claims] 1. A data sequence in which a run-length numerical sequence obtained by scanning data is sequentially input as a set of a continuous number of 0 bits and a continuous number of 1 bits as an input unit, converted into a predetermined code sequence, and output. An encoding method, wherein the data of the input unit is based on the number of consecutive 0 bits, the number of consecutive 1 bits, the number of consecutive 0 bits immediately before, and the number of consecutive 1 bits immediately before. A calculating step of calculating a corresponding state value, and a generating step of generating a state code and a numerical code corresponding to the data of the input unit based on the state value calculated by the calculating step and the immediately preceding state value. A data encoding method characterized by including: 2. The method according to claim 1, wherein the consecutive number X of the 0 bits, the consecutive number Y of the one bit, the consecutive number A of the immediately preceding 0 bit, and the consecutive number B of the immediately preceding 1 bit are X, ≠ A
When the relation of Y ≠ B is satisfied, the value of the state S0 is changed to the state value S.
When X = A and Y ≠ B, the value of the state S1 is set as the state value S. When X ≠ A and Y = B, the value of the state S2 is set as the state value S. 2. The data encoding method according to claim 1, wherein the value of the state S3 is set to the state value S when A and Y = B. 3. In the generating step, the immediately preceding state value T is the state S3, and the state value S
A repetition number calculating step of calculating the number N of repetitions in which the repetition event occurs successively when the coincidence event becomes equal to the immediately preceding state value T, and the occurrence of the repetition event A regression event determining step of determining a regression event when the status value S and the immediately preceding status value T differ from each other or when input is completed; and a value of the status value S when the mismatch event occurs. A state code generation step of generating a state code indicating: when the generation of the state code by the state code generation step is completed or when the coincidence event occurs, the continuous number X of the 0 bits or the 1 bit A first numerical code generation step of generating a numerical code indicating the value of the continuous number Y, and, when the regression event occurs, prior to generation of the state code by the state code generation step, the repetition number calculation step Yo The data encoding method according to claim 2, further comprising: a second numerical code generation step of generating a numerical code indicating the value of the number of repetitions N calculated in the above. 4. The calculating step includes, when coding is started, storing a predetermined initial value in each of T, A, B and the number of repetitions N, and a predetermined run length in the X and Y. An input step of inputting a value, and a state value generation step of generating the state value S after the input by the input step, wherein the generation step corresponds to the state value S generated by the state value generation step. A coincidence event determining step of determining whether the repetitive event has occurred; a repetitive event determining step of determining whether the repetitive event has occurred in response to the occurrence of the coincident event; A repetition number adding step of adding the number of repetitions, a regression event determination step of determining whether or not the regression event has occurred in response to the occurrence of the mismatch event or the end of the data to be encoded, To occur A number-of-repetitions number code generating step for generating a number-of-repetitions number code indicating the value of the number-of-repetitions N; a regression step of returning the number-of-repetitions N to a predetermined initial value after the generation of the number-of-repetitions number code; A state code generating step of generating a state code representing the value of the state value S in response to the occurrence of a mismatch event or the initialization of the number of repetitions N by the regression step; After the repetition event is determined by the repetition event determination step, a numerical code corresponding to X is generated when X does not match A, and a numerical code corresponding to Y when Y does not match B. After the numerical code generation step of generating the numerical code, and the generation of the numerical code by the numerical code generation step or the addition of the repetition number N by the repetition number addition step, the content of A is updated to X, and the content of B is changed to Y Update and before 4. The data encoding method according to claim 3, further comprising a step of updating immediately before the content of the description T to S. 5. A code string obtained by scanning code string data is input, restored to a run-length numerical string having a set of a number X of continuous bits of 0 and a number Y of continuous bits of 1 as a unit, and output. A data decoding method, comprising: a storage unit (S) for storing a state value to be restored when decoding is started; a storage unit (C) for storing a value corresponding to the X among the restored run-length values; An initialization step of substituting a predetermined initial value into a storage unit D for storing a value corresponding to the Y among the run-length values to be initialized, and inputting a predetermined code unit and setting the code to a numerical code P Or an input code type determining step of determining whether the code is a state code Q. When the input code type determining step determines that the code is a numerical code, the code is restored to a numerical value and stored in the storage unit L. A code restoring step, and the input code type determining step. A state code restoring step of restoring the code to the state value and storing it in the S when it is determined that the state code is a state code; A state S0 in a relation of ＳA and Y ≠ B, a state S1 in a relation of X = A and Y ≠ B, a state S2 in a relation of X ≠ A and Y = B, or a relation of X = A and Y = B (A is the number of consecutive bits of the immediately preceding 0, and B is the number of consecutive bits of the immediately preceding 1). At the time when the value is stored, an S3 state determination step of determining whether the value of the S matches the S3, and when the value of the S is identified as the S0 by the identification step, After storing the value of L in the C, the next numerical code is input and restored, and stored in the D. A first run value update step of storing the value of L in the D when the value of S is identified as S1 by the identification step; A third run value updating step of storing the value of L in the C when the value of S is identified as S2 by the identification step; and the first, second or third run. The value of C and D is output as a set of run-length values at the point in time when the update in the value update step is completed or when the value of S is determined to be equal to the state value S3 in the S3 state determination step. 1 and the number of times corresponding to the value of L as a set of run-length values when the value of S is identified to be S3 by the identifying step. The second to output
A data decoding method, comprising the steps of: 6. A computer-readable recording medium on which a program for causing a computer to execute the method according to claim 1 is recorded.